A Selective Cluster Index Scheduling Method in OFDMA

Abstract

Abstract-OFDM is an attractive solution for the design of future wireless communications due to its robustness to dispersion in multipath environments. Additional diversity gains can be realised by OFDMA by exploiting not only the temporal fading, but also spectral fading, that can result in higher rates due to increased diversity. A scheduler exploiting multiuser diversity imposes strict feedback requirements as channel quality information is required for every user's feedback unit with a feedback rate at least equal to the coherence time of the channel. The instantaneous signal to noise ratio is the most commonly used metric to quantify users' channel conditions. In OFDMA, the overhead due to feedback transmission increases significantly due to the fact that signal to noise ratio is required in the frequency domain, for every subcarrier of the OFDM symbol, which can potentially mitigate any multiuser diversity gains. This paper proposes a reduced feedback OFDMA design that requires very limited feedback information on a clustered subcarrier basis, whilst maintaining high downlink rates via Opportunistic Beamforming with multiple weighting vector transmission from the Base Station. This scheme is shown to be quite robust, as it can be easily adapted to variable channel conditions without any significant increase in complexity.

Reduced feedback selective cluster index scheduling with user pre-selection for next-generation multi-input multi-output orthogonal frequency division multiple access systems

“…Work in [20] has extended this concept, showing that considerable throughput benefits can be achieved via the combined use of multiple precoding and multiple threshold design which allows for a more efficient identification of strong users. The increase in uplink feedback overhead with respect to the original 1-bit feedback scheme has been justified by an improvement in the aggregate downlink throughput.…”

“…Beyond a certain cluster size however, S opt inevitably converges to unity, limiting the ability of the SCI scheme to achieve linear reductions in feedback. Nevertheless, even under such conditions, the proposed scheduling algorithm attains a 3-fold feedback saving from the least spectrally efficient, 1-bit per resource feedback approach, and up to a 7-fold increase over the more spectrally efficient multiple precoding scheme (Q=3) based on a two-threshold implementation [20]. The spectral efficiency for the SCI and the referenced threshold based schemes for K=50 is compared in Figure 7 (b).…”

“…Since the proposed metric provides only relative channel strength indication, this reduction in feedback allows for better selectivity, since non-preselected users are restricted from transmitting channel information for scheduling purposes. The proposed approach shows considerable uplink feedback savings over SNR based feedback schemes, as well as other existing reduced feedback schemes that do not rely on SNR feedback [8,20], since it allows for feedback savings via a reduction of the CQI size as well as reduction of the number of resources for which CQI is required.…”

Exploiting Multiuser Diversity in OFDMA Systems with Limited Feedback

“…Several algorithms were proposed in the literature to reduce the CSI feedback overhead. For instance, some of the proposed algorithms request users to feed back information for only a group of subcarriers called cluster [3], [13], [14]. The work in [3] proposed that each user feeds back the indices of the strongest clusters only.…”

“…For instance, some of the proposed algorithms request users to feed back information for only a group of subcarriers called cluster [3], [13], [14]. The work in [3] proposed that each user feeds back the indices of the strongest clusters only. The work in [4] and [5] introduced threshold based scheduling.…”